Spring Challenge 2020 - Bugs & Questions

TwoSteps · May 8, 2020, 7:27am

We’re looking into the server issues

_iftekhar · May 7, 2020, 8:03pm

Console is continuously showing error #73 when I click “play my code”.

BernardoSulzbach · May 7, 2020, 8:25pm

I have had the same issue for some time now, as I described in https://www.codingame.com/forum/t/spring-challenge-2020-bugs-questions/180397. I think this is happening to everyone.

HyperDec · May 8, 2020, 8:23am

I struggle with 504 or 74 error as well (and other errors about buffer). I work in Rust and it appears it was and infinite loop (shame on me).
This is not well handle by the referee, no timeout after 50 ms as usual with this. I was blocked but it load the server for nothing too.
Maybe a bug to look at for Rust and for other language ?

Bazuara · May 8, 2020, 9:29am

Hi everyone!
C code doesn´t work. When trying default code it takes over 2 minutes to execute, just to tell you that there was no input. If I switch to any other language, the response time is almost immediate and it responds to commands.

volrk · May 8, 2020, 9:49am

Hey

I don’t know if there is this in an other languge, but in rust when I get the input, if a line has no wall there is no space character.

here one exemple :

###################################
#   #     #   # # # #   #     #   #
# # # # ##### # # # # ##### # # # #
# #   #       #     #       #   # #
# ### ### # # # ### # # # ### ### #
# #   #   #             #   #   # #
# # # # # ### # ### # ### # # # # #
#   #   #     #     #     #   #   #
### ### ##### ### ### ##### ### ###

### # # # ##### ### ##### # # # ###
### # #     #         #     # # ###
### # ##### # # ### # # ##### # ###
###################################

the 10th row has no wall (’#’) but there is no space (’ ')
It is the same if a row does’t finish with a wall there isn’t the last spaces

HyperDec · May 8, 2020, 9:53am

Look at the inputs, there is a trim_end(), just delete that and it will be fine.

volrk · May 8, 2020, 9:54am

Thank, I didn’t see it

agysejt · May 8, 2020, 11:26am

Hello Community!

Default C# timed out in init message!

Stopwatch sw = new Stopwatch();
sw.Restart();
string[] inputs;
inputs = Console.ReadLine().Split(’ ');
Console.Error.WriteLine($“start input-1: {sw.ElapsedMilliseconds} msec”);

sw.ElapsedMilliseconds is higher than 3000 msec!!!
(it needs maximum 1 msec)

Could you be so kind to correct this problem?

Aquila13fr · May 8, 2020, 11:49am

Bonjour.
J’ai l’impression de mon classement n’evolue plus.
Dans le classement complet, je vois de la date de la derniere soumission n’est pas la bonne.
Est-ce normal ?

Hello.
I have the impression of my classification no longer evolving.
In the full ranking, I see the date of the last submission is not the right one.
Is this normal?

eulerscheZahl · May 8, 2020, 1:15pm

The problem is on your end: you are supposed to start the timer after reading the input. Otherwise you time the serverload too (starting the game, waiting for your opponent and what not).

GiB · May 8, 2020, 1:48pm

I don’t understand the interaction between a pac “crossing” the path of a stronger opponent pac:

github.com

CodinGame/SpringChallenge2020/blob/master/src/main/java/com/codingame/spring2020/Game.java#L633


    if (!Config.FRIENDLY_BODY_BLOCK && pac.getOwner() == other.getOwner()) {
        return false;
    }


    // Never blocked against something pac can eat
    if (canEat(pac, other)) {
        return false;
    }


    // If beaten, can go to same coord (we only block crossing in that case)
    if (canEat(other, pac) && pacmenWillShareSameCoord(pac, other)) {
        return false;
    }


    return pacmenWillCollide(pac, other);
}


private boolean pacmenWillShareSameCoord(Pacman a, Pacman b) {
    Coord toA = getIntendedPositionAtStep(a, a.gotBlocked() ? a.getCurrentPathStep() : a.getCurrentPathStep() + 1);
    Coord toB = getIntendedPositionAtStep(b, b.gotBlocked() ? b.getCurrentPathStep() : b.getCurrentPathStep() + 1);
    return toA.equals(toB);

Sample situation:
Blue speeded ROCK at (0,0)
Red speeded PAPER at (1,0)

Case 1:
Blue plays MOVE (2,0) (right-right), Red plays nothing
Does Blue get blocked or eaten ?

Case 1bis:
Blue plays MOVE (1,0) (right), Red plays nothing
Does Blue get blocked or eaten ?

Case 2:
Blue plays MOVE (0,2) (down-down), Red plays MOVE (0,1) (left-down)
Does Blue get blocked because the paths overlap at (0,0) and (0,1) ? Does it get eaten ?

Case 3:
Blue plays MOVE (0,2) (down-down), Red plays MOVE (-1,0) (left-left)
Same case, the paths overlap at Blue’s starting square

Case 4:
Blue plays MOVE (2,0) (right-right), Red plays MOVE (1,2) (down-down)
In this case, the paths overlap at Red’s starting square

Followup question: in cases 1 and 1bis, Red played nothing. Would playing MOVE (1,0) (staying on-place) be equivalent ?

GiB · May 8, 2020, 1:56pm

To summarize, I think this “crossing” special case is inconsistent: if a pac runs into a stronger pac, it should, not depending on the situation:

get blocked (-> return true in the code above)
or
suicide (-> remove the pacmenWillShareSameCoord condition in the code above)

I would favor (1) to promote chasing: if the opponent defends with a SWITCH, the chasing pac does not get killed and can flee next turn.

With (2) and the current rules, the chaser has to guess whether the defender will play a SWITCH or not.

DrBrask · May 8, 2020, 3:58pm

C# code is compiled in debug mode, without optimization and with debug tracking enabled.
Also unsafe code is not enabled…

agysejt · May 8, 2020, 5:04pm

No! It is debug compilation… Try the default code, in some case will be timed out…

StrafeAE · May 8, 2020, 11:26pm

Quick question: Can a pac see pellets on the other side of the map through tunnels?

programjames · May 8, 2020, 11:59pm

Where do you go to see school/company rankings? I have added my school to my profile but I can’t find its rank.

GameRaccoon · May 9, 2020, 7:46am

Is “speed up” broken or it just me?
If I test in the IDE then only one team got affected by speed up:

if both teams always use speed up when they can, then only one of the teams moves twice as fast during the speed up
if only one team uses speed up, then sometimes speed up has no effect at all

I thought it is something with the IDE testing, but just noticed that in the battles doesn’t seems to work properly either.

Example 1:

Note that at the very first turn the red team uses speed up and the blue team does not (at least there’s no visual feedback for that), and both teams move at the same speed after this.

Example 2 shows that it’s not a visual bug.

At step 41 two pacs of the blue team use SPEED and the third uses SWITCH. After that all three pacs of the blue team move at the same speed.

Fangel · May 9, 2020, 10:19am

The fact that the fog of war is limited to horizontal and vertical line of sights, it implies 2 things:

2 opponents pacs may cross without having seen each other before. And then, it is mainly up to luck. Of course you can try to predict the opponent moves, but from a pac that disappeared from radar 10 turns ago, it is complicated. It may have wait, speed up or even transformed to any type. Moreover, you cannot play safely as these dangerous crossing points are everywhere.
you may try to eat an opponent but you cannot be sure of what happened as it could have left your line of sight.

I would suggest then to list the opponents pacs that are still alive, without giving position, state or features.
Another suggestion would be to give sight on direct diagonals to prevent random encounters and gives the game more strategy. One would still be able to ambush properly using a speed.

RehanRRajput · May 9, 2020, 11:24am

Hi,

In this replay: https://www.codingame.com/replay/459438636

At step 14, the game is over even though I have at least 3 Pacs outside the line of sight of the opponent (they can’t kill my pacs in the next step).

Is it a bug or there’s something that I missed in the instructions?